The Pedagogy Is Built In, Not Bolted On
Most AI course generators produce a summary with a quiz stapled to the end. AmpUp reads the source material and the performance goal, infers which instructional design model actually fits, and structures the course around it. You can see the model it chose, change it, or pin the one your team standardizes on.
ADDIE
Analyze, Design, Develop, Implement, EvaluateThe default spine for a full program. AmpUp runs the analysis against your source corpus and your performance data, proposes objectives and a module structure, generates the build, and wires the evaluation in before launch instead of after it.
Bloom's Taxonomy
Remember through CreateEvery objective is written at an explicit cognitive level, and the assessment is generated to match. Recall objectives get quizzes. Apply and evaluate objectives get a live roleplay or a walkthrough, because a multiple-choice question cannot test them.
Gagné's Nine Events
Instructional sequencingLesson-level structure: gain attention, state the objective, recall prior learning, present the material, guide, elicit performance, give feedback, assess, and support transfer. This is why an AmpUp lesson opens and closes the way it does.
Kirkpatrick's Four Levels
Reaction, Learning, Behavior, ResultsMost training tools stop at level one. AmpUp instruments all four: reaction from engagement events, learning from assessment and rubric scores, behavior from what changes on real recorded conversations, and results from the deal outcomes those conversations produce.
Merrill's First Principles
Problem-centered instructionCourses are built around a real task rather than a topic list: activate what the learner already knows, demonstrate it, have them do it, then integrate it into the job. The roleplay checkpoint is the demonstration and application step, not a bonus feature.
SAM and rapid iteration
Successive approximationFor teams that iterate rather than waterfall, AmpUp fits the SAM loop: a reviewable prototype early, plain-English feedback, and a regenerated version in the same sitting instead of the next sprint.
Learning objectives are written at a stated Bloom's level, and the assessment is generated to match it. A course that claims “evaluate” does not get to end in a recall quiz.